CCminer(SP-MOD) Modded GPU kernels.

[scryptr]

@scryptr its for blake-256 ie blakecoin algo.

BLAKECOIN--

Thanks.  I did some checking.  BlakeCoin spiked yesterday.  Apparently, an FPGA can process the algo, a 750ti is OK, and a 980ti will match an AMD 290X.

--scryptr

[bathrobehero]

@impulse2000 what does your bat file look like for myr-gr algo? does the pool-side hashrate agree with your local hashrate? 
i'm still only able to get 10% of local hashrate to show at the pool , even using --diff 0.039xxxx

If the hashrate is too low poolside, increase the diff until you get rejects locally.

[pallas]

End of the game for neoscrypt as well:

https://bitcointalk.org/index.php?topic=632503.msg13373431#msg13373431

[hashbrown9000]

End of the game for neoscrypt as well:

https://bitcointalk.org/index.php?topic=632503.msg13373431#msg13373431

even at 10 mBTC/GH/day neoscrypt was just barely profitable for maxwell.  and it hasn't seen that rate in a while

[hashbrown9000]

If the hashrate is too low poolside, increase the diff until you get rejects locally.

--diff 0.00390625 is the only setting that gets me no local rejects but only gives 1/10 the hash on the pool.  All other diff settings in multiples of 2 or divisions of 2 give rejects and still not the correct hash at the pool.

if i increase the diff over and above 1.0 in multiples of 2, the miner starts submitting faster and faster with 0 accepts.

[djm34]

End of the game for neoscrypt as well:

https://bitcointalk.org/index.php?topic=632503.msg13373431#msg13373431

300kh/s isn't really impressive  my version is doing 850kh on the 980

[sp_]

300%

A r9 280x did 200khash with the old kernal. and 600 now?

[sp_]

@scryptr its for blake-256 ie blakecoin algo.

BLAKECOIN--

Thanks.  I did some checking.  BlakeCoin spiked yesterday.  Apparently, an FPGA can process the algo, a 750ti is OK, and a 980ti will match an AMD 290X.

--scryptr

Blakecoin is faster compiled with cuda 7.5 (latest release 77). What are your numbers?

[scryptr]

@scryptr its for blake-256 ie blakecoin algo.

BLAKECOIN--

Thanks.  I did some checking.  BlakeCoin spiked yesterday.  Apparently, an FPGA can process the algo, a 750ti is OK, and a 980ti will match an AMD 290X.

--scryptr

Blakecoin is faster compiled with cuda 7.5 (latest release 77). What are your numbers?

NEOSCRYPT--

I haven't switched to dot 77+ yet because of Neoscrypt.  I haven't mined any Blake algo coins either, but apparently BlakeCoin was mineable with CudaMiner back in 2013-2014.

My GTX 960 gets 340kh/s mining Neoscrypt with the latest NiceHash Miner, that makes my GTX 960 as fast as an AMD 280X with the latest NiceHash Miner.  They are still using your release dot 74.  If you did a hybrid compile (Pallas), NiceHash may incorporate your release dot 77+.  Neoscrypt is still a top-paying algo at NiceHash, and AMD code is really being optimized there.

I may do some benchmark tests.       --scryptr

[sp_]

@scryptr its for blake-256 ie blakecoin algo.

BLAKECOIN--
Thanks.  I did some checking.  BlakeCoin spiked yesterday.  Apparently, an FPGA can process the algo, a 750ti is OK, and a 980ti will match an AMD 290X.
--scryptr

Blakecoin is faster compiled with cuda 7.5 (latest release 77). What are your numbers?

NEOSCRYPT--
I haven't switched to dot 77+ yet because of Neoscrypt.  I haven't mined any Blake algo coins either, but apparently BlakeCoin was mineable with CudaMiner back in 2013-2014.
My GTX 960 gets 340kh/s mining Neoscrypt with the latest NiceHash Miner, that makes my GTX 960 as fast as an AMD 280X with the latest NiceHash Miner.  They are still using your release dot 74.  If you did a hybrid compile (Pallas), NiceHash may incorporate your release dot 77+.  Neoscrypt is still a top-paying algo at NiceHash, and AMD code is really being optimized there.
I may do some benchmark tests.       --scryptr

Blakecoin is minable in cudaminer, but ccminer sp-mod release 77 is 100% faster.

So 340Khash @ r9 280x is only 60% bether than the opensource.. But on the Nano and the furyx the rate is above 600khash.

[sp_]

End of the game for neoscrypt as well:
https://bitcointalk.org/index.php?topic=632503.msg13373431#msg13373431

No point in mining Neoscrypt on AMD Cards when you can mine etherum and double the profit.

[pallas]

End of the game for neoscrypt as well:
https://bitcointalk.org/index.php?topic=632503.msg13373431#msg13373431

No point in mining Neoscrypt on AMD Cards when you can mine etherum and double the profit.

I was just saying that we (as nvidia) don't have the edge on neoscypt any longer.
And that the payouts will get lower 'cause of the higher hashrate.

[djm34]

End of the game for neoscrypt as well:
https://bitcointalk.org/index.php?topic=632503.msg13373431#msg13373431

No point in mining Neoscrypt on AMD Cards when you can mine etherum and double the profit.

I was just saying that we (as nvidia) don't have the edge on neoscypt any longer.
And that the payouts will get lower 'cause of the higher hashrate.

still have 

[pallas]

End of the game for neoscrypt as well:
https://bitcointalk.org/index.php?topic=632503.msg13373431#msg13373431

No point in mining Neoscrypt on AMD Cards when you can mine etherum and double the profit.

I was just saying that we (as nvidia) don't have the edge on neoscypt any longer.
And that the payouts will get lower 'cause of the higher hashrate.

still have 

YOU still have :-D

[djm34]

End of the game for neoscrypt as well:
https://bitcointalk.org/index.php?topic=632503.msg13373431#msg13373431

No point in mining just donate.

[sp_]

Let me guess..

Shavite uint4 rewrite with loops to reduce the codesize?

[sp_]

nice. Want to share the code?

[sp_]

nice. Want to share the code?

I don't give away answers, you know that. But hints or exchanging ideas I'm always happy to do. There may be a better way to improve Echo than what I've done, actually, however - it will lend itself to Nvidia's architecture more - but at the cost of a slightly more complex round. Best to code for your platform, though... I might give it a go..

The bottleneck in the echo is the table based AES. But it might help to split the work in the kernal into two passes to improve the register count.
Compiling with cuda 7.5 causes a register and stack explosion.

[sp_]

nice. Want to share the code?

I don't give away answers, you know that. But hints or exchanging ideas I'm always happy to do. There may be a better way to improve Echo than what I've done, actually, however - it will lend itself to Nvidia's architecture more - but at the cost of a slightly more complex round. Best to code for your platform, though... I might give it a go..

You should pickup a couple of cheap and used 750ti's on ebay. I have seen them below $100.

[sp_]

But I have another plan. Reduce the table accesses in the AES by 50% by doing lookups on 12 bits instead of 8bits.
You need 16KB of shared mem instead of 1kb. but still under the 96kb limit on the compute5.2 cards and 64kb limit on the compute 5.0
Then you could do AES with 50% less instructions and 50% less shared mem access..

But you need to rewrite/redesign the algorithm. Alot of work.